Engineering a Carbonyl Reductase for Scalable Preparation of (S)-3-Cyclopentyl-3-hydroxypropanenitrile, the Key Building Block of Ruxolitinib

Graphical Rational engineering of the carbonyl reductase PhADH resulted in a double mutant H93C/A139L, which enhanced the enantioselectivity toward the reduction of 3-cyclopentyl-3-ketopropanenitrile 1 a from 85 % to 99 %, as well as a 6-fold improvement in the specific activity. ( S )-3-Cyclopentyl-3-hydroxypropanenitrile (( S )- 1 b ) was prepared at 200 g/L (1.5 M) substrate concentration. ( S )-3-Cyclopentyl-3-hydroxypropanenitrile is the key precursor for the synthesis of ruxolitinib. The bioreduction of 3-cyclopentyl-3-ketopropanenitrile ( 1 a ) offers an attractive method to access this important compound. A carbonyl reductase (PhADH) from Paraburkholderia hospita catalyzed the reduction of 1 a giving the ( S )-alcohol ( 1 b ) with 85 % ee. Rational engineering of PhADH resulted in a double mutant H93C/A139L, which enhanced the enantioselectivity from 85 % to >98 %, as well as a 6.3-fold improvement in the specific activity. The bioreduction of 1 a was performed at 200 g/L (1.5 M) substrate concentration, leading to isolation of ( S )- 1 b in 91 % yield. Similarly, using this mutant enzyme, 3-cyclohexyl-3-ketopropanenitrile ( 2 a ) and 3-phenyl-3-ketopropanenitrile ( 3 a ) were reduced at high concentration affording the corresponding alcohols in >99 % ee, and 90 % and 92 % yield, respectively. The results showed that the variant H93C/A139L was a powerful biocatalyst for reduction of β-substituted-β-ketonitriles.